Conversational speech datasets: 59 listings
Spirelight publishes 59 listings for existing conversational speech recordings by language, dialect and region. Request a sample to check recording format and transcript coverage. Ask our team for provenance and consent documentation, then confirm availability and written licence terms for your intended use. Our guide to buying AI training data explains how to compare samples and set acceptance checks.
Need a custom speech dataset? If the listed language, speaker mix, device or recording conditions do not fit your project, ask us to scope a new collection. Custom work is assessed and quoted separately.
In stock now, with download access within 24 hours of a signed order: American English, 965 hours, French, 1,131 hours and German, 1,534 hours.
Listing metadata date: . The listings declare 71,130 audio hours in total, with listed prices of 35 to 95 USD per audio hour. Confirm usable hours, sample coverage and the applicable terms for the specific asset you need. For a requirement outside the catalogue, describe your brief so we can assess custom speech data collection feasibility, scope and schedule.
The video couldn’t start. Try playing it again or open the video directly.
Showing 7 of 59 datasets
Arabic (Modern Standard)
Saudi Arabia · ConversationalArabic (Gulf)
Saudi Arabia · ConversationalArabic (Levantine)
Lebanon · ConversationalArabic (Egyptian)
Egypt · ConversationalArabic (Darija)
Morocco · ConversationalBengali
Bangladesh · ConversationalBulgarian
Bulgaria · ConversationalCatalan
Spain · ConversationalCroatian
Croatia · ConversationalCzech
Czechia · ConversationalDanish
Denmark · ConversationalDutch
Netherlands · ConversationalDutch (Surinamese)
Suriname · ConversationalEnglish (US)
United States · In stockEnglish (South African)
South Africa · ConversationalEnglish (Singaporean)
Singapore · ConversationalFinnish
Finland · ConversationalFrench
Francophone Africa · In stockGerman
Germany · In stockGreek
Greece · ConversationalGujarati
India · ConversationalHausa
Nigeria · ConversationalHebrew
Israel · ConversationalHindi
India · ConversationalHungarian
Hungary · ConversationalIndonesian
Indonesia · ConversationalItalian
Italy · ConversationalJapanese
Japan · ConversationalKannada
India · ConversationalKorean
South Korea · ConversationalMalay
Malaysia · ConversationalMalayalam
India · ConversationalMandarin Chinese (Taiwanese)
Taiwan · ConversationalMarathi
India · ConversationalNorwegian
Norway · ConversationalPersian (Farsi)
Iran · ConversationalPolish
Poland · ConversationalPortuguese (European)
Portugal · ConversationalPortuguese (Brazilian)
Brazil · ConversationalPortuguese (African)
Angola · ConversationalPunjabi
India · ConversationalRomanian
Romania · ConversationalRussian
Russia · ConversationalSerbian
Serbia · ConversationalSlovak
Slovakia · ConversationalSpanish (European)
Spain · ConversationalSpanish (Latin American)
Mexico · ConversationalSwahili
Kenya · ConversationalSwedish
Sweden · ConversationalTagalog
Philippines · ConversationalTamil
India · ConversationalTelugu
India · ConversationalThai
Thailand · ConversationalTurkish
Turkey · ConversationalUkrainian
Ukraine · ConversationalUrdu
Pakistan · ConversationalVietnamese
Vietnam · ConversationalYoruba
Nigeria · ConversationalNo dataset matches that search. Describe the language, recording conditions and intended use so we can assess custom speech data collection feasibility and scope.